• No results found

Credit rating agencies have been repeatedly blamed for causing or exacerbating negative market developments in the context of the European sovereign debt crisis1. Some recent developments, in particular the erroneous announcement by Standard & Poor’s (S&P) to some of its clients of a downgrade of France, have added to volatility.

However, the question of the extent to which credit ratings exacerbate fluctuations in sovereign credit markets is far from trivial. Consensus on sovereign bond markets often evolves faster than credit ratings, which tend to typically be “behind the curve”. Negative ratings decisions are often made after the deterioration of market-based credit indicators. This has been a consistent pattern since the start of the financial crisis. In that sense, ratings can be considered as lagging indicator that often show only information that is already known by the market.

When negative ratings decisions are made, they are unsurprisingly generally associated with yield increases (an outlier was the downgrade of the United States by S&P in early August 2011, which was associated with a general increase in risk aversion and lower yields on US bonds because of their safe-haven status). This effect is confirmed by recent studies such as those published by the International Monetary Fund (Arezki, Candelon & Sy, 2011) or the European Central Bank (Afonso, Furceri & Gomes, 2011). However, the extent of this impact is less clear. In particular, the ECB study notes that negative ratings decisions tend to be preceded by negative market developments, raising questions about the direction of causality. Consistent with previous literature, the ECB paper confirms the existence of a significant reaction, on the part of both sovereign yields and CDS spreads, to rating announcements (this is true in particular for negative events). The analysis goes a step further since it assesses if both sovereign yields and CDS spreads had already absorbed the information contained in changes to ratings before their announcement. As regards the anticipation mechanism, the main result is that the information contained in rating announcements is not anticipated by the credit market while the CDS market seems to anticipate the information contained in rating downgrades. The ECB study specifically investigated the issue of causality between ratings changes and yields/CDS spread over the short-term and concluded that there is “two-way causality between sovereign credit ratings and government bond yield spreads”, namely that “past values of changes in yield (CDS) spreads are significant determinants of the change in effective rating and vice-versa” (Afonso, Furceri & Gomes, 2011).

Furthermore, this is not a static picture as investors’ behaviour changes over time. Anecdotal evidence suggests a gradually reduced dependence on credit ratings since the start of the euro-area crisis. For example, some large investors appear to have moved away from relying on ratings-based sovereign-bond benchmarks indices to form their own benchmarks with no reliance on ratings. Strikingly, some negative ratings decisions during the past 18 months have had negligible market impact, suggesting that many commentators’ emphasis on “mechanical effects” does not capture the complexity of

____________________________________________________________________________________________  especially after the G20 Deauville declaration of 18 October 2010, market developments appeared to be driven more by political pronouncements and less by ratings decisions. The Bank for International Settlements has concluded that the Deauville declaration had significant market impact (BIS, 2010).

Rating Agencies and Sovereign Credit Risk Assessment

____________________________________________________________________________________________

2.

REDUCING OVER-RELIANCE ON CREDIT RATINGS

Credit rating agencies derive some of their importance from the fact that the regulatory

s are indeed undesirable,

t is to be expected that ratings will be complemented with other

full implementation of the system relies on their assessments. This reliance is observed in bank regulation, which in some circumstances sets banks’ capital requirements in relation to asset risks as assessed by CRAs. Similar regulations exist for insurance and other financial market participants. Following the failures of ratings in the US sub-prime mortgage-based securities market, significant work has been undertaken by regulators and supervisors, at the global level and on both sides of the Atlantic, to reduce regulatory reliance on credit ratings. The most radical initiative so far has been the decision by the US Congress to ask federal supervisors to eliminate all references to credit ratings in their rules (Section 939A of the US Dodd- Frank Act of July 2010). However, implementing this decision is proving difficult, not least because it impedes the adoption by the US of global supervisory standards (“Basel 2.5” and Basel III) which do refer to credit ratings2. At the global level, a review of this issue by the Financial Stability Board has concluded that “in certain cases, it may take a number of years for market participants to develop enhanced risk management capability so as to enable reduced reliance on credit rating agencies” (FSB, 2010).

One problem is that, while references to risk ratings in regulation

the alternatives might be even worse. In particular, banks’ own models of risk assessment have been proven in the crisis to be even less reliable than credit ratings, including in the largest banks where risk management was widely believed to be most advanced (see for example UBS, 2008). Replacing references to ratings with references to market-based risk indicators may sharply increase pro-cyclicality, as such indicators are typically much more volatile than credit ratings (see for example Moody’s, 2009, in the case of corporate ratings).

As a consequence, i

measures of risk, but that a complete elimination of references to credit ratings from the European financial rulebook would appear both impractical and undesirable given the lack of proper alternatives in many cases. Moreover, contemplating such steps in the current period of market stress may contribute to short-term volatility.

In particular, the EU and its member states should proceed with

“Basel 2.5” and Basel III accords, including the extent to which these still refer to the use of credit ratings in spite of the Basel Committee’s efforts to reduce reliance. These efforts by the Basel Committee and other international financial standards-setters are expected to continue and to bring gradual improvements in the years to come. Opting for a complete elimination of any regulatory reference to credit ratings in the short term would have significant downsides.

____________________________________________________________________________________________ 

3.

A EUROPEAN RATING FOUNDATION

In a resolution adopted in June 2011, the European Parliament asked the European Commission to study the creation of a new fully-independent European Credit Rating Foundation (European Parliament, 2011). The resolution does not include an explicit deadline for this. The Commission's proposal for a third EU Regulation of CRAs (CRA3) does not retain the option of EU-level public sponsorship of a new CRA: “This proposal is not aimed at setting up a European credit rating agency. As requested by the European Parliament in its report on credit rating agencies of 8 June 2011, this option was assessed in detail in the impact assessment accompanying this proposal. The impact assessment found that even if a publicly funded CRA may have some benefits it terms of increasing the diversity of opinions in the rating market and providing an alternative to the issuer pays model, it would be difficult to address concerns relating to conflicts of interest and its credibility, especially if such CRA would rate sovereign debt. However, these findings should by no means discourage other actors from setting up new credit rating agencies. The Commission will monitor to what extent new private entrants in the credit rating market will provide for more diversity” (European Commission, 2011). In June, the Parliament also proposed to establish a European rating index (EURIX), incorporating all ratings of registered CRAs that are available on the market (European Parliament, 2011). Indeed, while more competition in the credit ratings market is desirable, it is not clear that this can be achieved through a public initiative. A publicly-sponsored ratings agency would be assumed by market participants to be politically constrained in its credit assessments and would therefore struggle to make a difference in terms of market perceptions – especially in a context in which there is a widespread perception that EU authorities are tempted to increase political leverage over ratings decisions generally, as illustrated by the debate on the preparation of the CRA 3 Regulation before the publication of the European Commission’s proposal in mid-November 2011.

Incidentally, it is not clear that a specialisation in sovereign ratings could represent a sustainable business model for a financially independent ratings agency. Sovereign ratings by the three most established CRAs do not generate significant revenue3. CRAs rate the largest sovereigns not as a direct revenue generator, but because it is a necessary building block for other, more lucrative ratings segments such as those of corporate issuers. Indeed, many sovereign ratings are unsolicited, especially for the largest sovereign issuers, and as a consequence they are a pure cost centre for the CRAs. The low financial dependence on sovereign ratings also results in less obvious conflicts of interest in rating sovereigns than in other segments of CRA activity.

Rating Agencies and Sovereign Credit Risk Assessment

____________________________________________________________________________________________

4.

ASSUMPTION OF AN EXPLICIT RATING ROLE BY